U.S. Department of Labor, Bureau of Labor Statistics
Statistical Analysis Research and Computational Modeling
Small Area Estimates of Employment
Scope of Work: Developing a statistically sound method of small area estimation for employment. The method used relies on cross-sectional information in past censuses that is relevant to small domains defined by intersection of Metropolitan Statistical Area and Super Sectors. The work involves conducting statistical fitting of data including seasonality; determining 'best' overall estimator based on estimates of variances and correlations; and developing, modifying, and running the previously available ARIMA models.
Local Area Unemployment Statistical Models
Improving a benchmarking procedure, state-space models for unemployment estimates, which are currently fitted independently between states and combine a model of population values with that of sampling errors. At the end of each calendar year, the model dependent estimates are modified to guarantee that the annual mean estimate equals the corresponding mean sample estimate. The proposed solution under study is to fit the model jointly to several 'homogeneous states' with the added constraints
Results: Help establish better estimation techniques leading to fewer errors in the projection of unemployment rates in smaller areas
Prime: CIPS |
